01
Frontend
HTMLCSSJavaScriptTypeScriptReactNext.jsTailwind CSS
I build tools I want to use.
Backend and RAG for LLM agents. Warsaw.
Mostly backend: data models, what an agent actually gets in its context, and where the boundaries between services go. Right now I'm building Ariadne, a memory layer for LLM agents.
Skills
From the data layer up to the interface. I build whole systems, not single pieces.
02
UI/UX
FigmaAdobe Photoshopinterface designresponsivenessUX basics
03
Backend & data
Node.jsHonoAPI routesDrizzlePrismaPostgreSQLpgvectorrow-level securityZodbasic Python
04
Tools & workflow
GitGitHubDockerVercelMCP serversembeddingsClaude Codeworking with documentation
05
Soft skills
Fast learningcommunicationcreativityattention to detailteamworkwillingness to grow
Featured projects
Selected work. The rest is on GitHub.
AI Infrastructure
01Ariadne
Memory layer for LLM agents. The repo says what the code does, Ariadne holds why it looks that way.
TypeScriptMCPpgvectorTauri
284 commits30 branches
├─chore/clean-repo
├─docs/readme-project-story
└─feature/agent-marks
+27 more
$ git log --oneline
65ffb03docs: point the README at the 0.1.1 installer24d ago
▊
Fullstack Mobile App02
Ritatu
A macro tracker for Android I built after getting fed up with apps like Fitatu, overpriced and the AI barely works.
93 commits14
$ git log --oneline
e939445Merge pull request #16 from r1zuuu/feature/favorites9d ago
React NativeExpoTypeScript
View details
Hello!
I'm Stanisław. I build systems and learn as I go.
What pulls me in is RAG and the layer under the agent: data models, choosing what makes it into the context, and how the system splits into services. Lately that means memory: what gets stored, how you find the right piece of it later, and who is allowed to read it. I know how LLMs work under the hood, and I like algorithms, because there something is either correct or it isn't. Still learning, but the projects work and I own every decision in them.
+
Years learning
+
Projects
+
Hackathons
+
GitHub commits
Let's
talk.
I build backends for LLM agents: Postgres with pgvector, memory, MCP servers. If that's the kind of thing you're working on, write.